JP3436742B2 - Splitterless multi-carrier modem - Google Patents

Description

Detailed Description of the Invention

CROSS REFERENCE TO RELATED APPLICATIONS This application is based in part on the following applications filed by one or more of the inventors of the present application.

[Spl] filed October 10, 1997
itterless Multicarrier Mo
duration For High Speed D
ata Transport Over teleph
Richard G, entitled "One Wires"
Ross, John Greszcuk, DaveKr
insky, Marcos Tzannes, and M
United States Provisional Patent Application Serial Number 60 / 061,689 of Ichael Tzannes, January 16, 1998
“Dual Rate Multicarr”
ier Transmission System I
n A Splitterless Configur
Richard Gross, entitled "ation"
And Michael Tzannes, US Provisional Patent Application Serial Number ***, filed January 21, 1998, "Dual Rate Multicarrier."
rTransmission System In
A Splitterless Configurat
Richard Gross, M, entitled "ion."
arcos Tzannes, and Michael
Tzannes US provisional patent application serial number ***
No., filed Jan. 26, 1998, “Multical
rier SystemWith Dynamic P
Richards, entitled "Power Levels"
US provisional patent application serial number *** from Gross, Marcos Tzannes, and Michael Tzannes.

The entire disclosure of the above application is incorporated herein by reference.

BACKGROUND OF THE INVENTION A. FIELD OF THE INVENTION The present invention relates to telephone communication systems, and more particularly to telephone communication systems that use discrete multitone modulation to transmit data over digital subscriber lines.

B. The prior art public switched telephone network (PSTN) provides most individuals and businesses with the most widely available form of electronic communication. Due to the ready availability of PSTN and the considerable cost of providing alternative equipment, PSTN has
There is an ever-increasing demand for accepting the growing demand for transmitting significant amounts of data at high rates.
PTSN was originally configured to provide voice communication,
As a result, they have narrow bandwidth requirements and are increasingly relying on digital systems to meet service demands.

A major limiting factor in the ability to achieve high rate digital transmission has been the subscriber loop between the telephone central office (CO) and the subscriber's premises. This loop most commonly contains a single twisted wire pair. This twisted pair is 0
It is very suitable for carrying low frequency voice communications, where a bandwidth of ~ 4 kHz is quite suitable, but for wideband communications (ie, bands on the order of hundreds of kilohertz or more, unless new technologies for communications are adopted). Width) does not easily adapt.

One approach to this problem is
Discrete Multitone Digital Subscriber Line (DMT DS
L) technology and its variants, the discrete wavelet multitone digital subscriber line (DWMT DSL)
It was the development of technology. In the following, these and other forms of discrete multitone digital subscriber line technology (ADSL, HDSL, etc.) are generally referred to as "DSL technology" or often simply "DSL". The operation of the discrete multitone system and the application of the discrete multitone system to the DSL technology are described in "Multicarr".
ier Modulation For Data T
ranmission: An Idea Whos
e Time Has Come ", IEEE Com
communications Magazine, May,
1990, pp. More fully described in 5-14.

In DSL technology, communication between the central office and the subscriber's premises via a local subscriber loop is accomplished by modulating the data transmitted on a number of discrete frequency carriers. These discrete frequency carriers are grouped together and then transmitted via the subscriber loop. These carriers individually form discrete non-overlapping communication subchannels of limited bandwidth, and collectively, these carriers form what is in effect a wideband communication channel. Carriers are demodulated at the receiver end and data is recovered from these carriers.

The data symbol transmitted over each sub-channel has a number of bits which may vary from sub-channel to sub-channel depending on the signal-to-noise ratio (SNR) of the sub-channel. The number of bits that can be achieved under certain communication conditions is known as the "bit allocation" of a subchannel, and each subchannel as a function of the measured SNR of the subchannel and the bit error rate associated with this SNR. The channel is calculated in a known manner.

The SNR of each subchannel is determined by transmitting a reference signal over the various subchannels and measuring the SNR of the received signal. The loading information is typically the receiving or "local" end of the subscriber line (ie, the subscriber's home in the case of transmission from the central office to the subscriber, from the subscriber's home to the central office). In the case of transmission, it is calculated at the central office) and transmitted to the other (transmitting or "remote") end, so that each transmitter-receiver pair communicating with each other communicates with the same information. Bit allocation information is stored at both ends of the communication pair link and is used to define the number of bits used on each sub-channel when transmitting data to a particular receiver. Other subchannel parameters such as subchannel gains, time and frequency domain equalizer coefficients, and other characteristics may also be stored to help define the subchannels.

Of course, information can be transmitted in either direction over the subscriber line. For many applications, such as delivery of video to subscribers, delivery of internet services, etc., the bandwidth required from the central office to the subscriber is many times the bandwidth required from the subscriber to the central office. Bandwidth. One recently developed service that provides such capability is based on multi-tone asymmetric digital subscriber line (DMT ADSL) technology.
In one form of this service, each subchannel has 43
Up to 256 sub-channels with a bandwidth of 12.5 Hz are dedicated for downstream communication (from the central office to the subscriber's home), with each sub-channel also having 4312.5 Hz.
Up to 32 sub-channels, which has a bandwidth of 1 to provide upstream (subscriber home to central office) communications. Communication is by "frames" of data and control information. A
In the currently used form of DSL communication, 68 data frames and one sync frame form a "super frame" that is repeated during transmission. The data frame has data to be transmitted. Synchronous or "sync" frames are well-known bit sequences used to synchronize transmitting and receiving modems, further facilitating determination of transmission subchannel characteristics such as signal-to-noise ratio ("SNR"), among others. Provides a bit sequence to

[0012] Such systems, in fact, provide significantly increased bandwidth for data communications, but may occur over the subscriber line at the same time as wideband data is being carried. Special precautions are required to avoid interference with certain common voice communications and associated signaling and interference from this common voice communications and associated signaling. The signaling activity is
In general, includes transmission of, for example, ringing signals, busy tones, off-hook indications, on-hook indications, dialing signals, and the like, and associated actions such as picking up a phone off-hook, returning a phone on-hook, dialing, etc. . These voice communications, and the signaling associated with voice communications, are commonly referred to as "plain old telephone service" or POTS, by modulating data communications to frequencies higher than those currently used for POTS, Separated from data communication. The data communication and POTS signals are then separately retrieved by appropriate demodulation and filtering. A filter that separates data communication and POTS is generally “PO
It is called "TS splitter".

Voice and data communications must be separated at both the central office and the subscriber's premises, and thus P
The OTS splitter must be installed in both places. Central office installations are generally not a major issue because a single modem at the central office can handle multiple subscribers, and technicians are usually available at the central office. Installation in the customer's house is a problem. Typically, a trained technician must visit the homes of all subscribers who want to use this technology to make the required installations. In this regard, depending on the desired location of the ADSL device, extensive rewiring may have to be done. This is expensive and precludes widespread use of DSL technology.

The DSL system also includes other data services (ADSL, HDSL, I) on adjacent telephone lines.
Interference from SDN or TI services, etc.). These services may be started after the target ADSL service has already been started and the DSL for internet access is imagined as being in service at all times, so the effects of these disturbances are improved by the target ADSL transceiver. There must be.

(Summary of the Invention) A. OBJECTS OF THE INVENTION Accordingly, it is an object of the present invention to provide an improved digital subscriber line communication system.

It is a further object of the present invention to provide a digital subscriber line communication system that is compatible with existing voice communication services and does not require the use of POTS splitters.

Another object of the present invention is to efficiently handle data communications despite random interruptions associated with the simultaneous carrying of voice communications, or interference resulting from simultaneous data services on adjacent telephone lines. An object is to provide an improved digital subscriber line communication system.

B. SUMMARY OF THE INVENTION Splitterless Operation The invention described herein uses discrete multitone digital technology (DMT) over a digital subscriber line (DSL) in the presence of voice communications and other interference. In a system for communicating data, it is directed towards increasing the accuracy and reliability of communication. For ease of reference, the apparatus and method of the present invention are hereinafter collectively referred to simply as a modem. One such modem, typically located in a customer's premises, such as a home or office, is "downstream" from a central office with which the modem communicates. The other is typically located at the central office and is "upstream" from the customer's home. In line with industry practice, modems are often referred to herein as "ATU-R"("ADSL transceiver unit, remote", ie, located in a customer's premises).
And "ATU-C"("ADSL transceiver unit, central office"). Each modem includes a transmitter section for transmitting data and a receiving section for receiving data, and each modem is a discrete multitone type. That is, each modem transmits data over multiple subchannels of limited bandwidth. Typically,
The upstream or ATU-C modem transmits data to the downstream or ATU-R modem via a first set of subchannels, which are generally higher frequency subchannels, and is typically lower subchannel. Via a second, usually smaller, set of channels, downstream or AT
Receives data from the U-R modem.

Heretofore, such modems, when used on lines carrying both voice and data, have P
Needed an OTS splitter. According to the invention,
Applicants provide a data modem that carries voice and data communications simultaneously and that is used in a discrete multitone communication system that operates without the special filtering provided by POTS splitters. Therefore, these are "splitterless" modems. In the absence of certain disturbances, referred to herein as "disturbance events" and described more fully below, the modem of the present invention is determined by the transmission capabilities of the system without considering such disturbances. Transmit data at a rate. Preferably, this is the maximum data rate that can be provided for a particular communication sub-channel subject to predefined constraints such as maximum bit error rate, maximum signal power, etc. that can be imposed by other considerations. However, when a jamming event occurs on the communication channel, the modem of the present invention detects the event and immediately modifies subsequent communication operations. In other responses, the modem changes the bit allocation (and thus possibly the corresponding bit rate) between the sub-channels and the sub-channel gain to interfere with voice communication activity and from voice communication activity. Limit the interference or compensate for interference from other services or sources that are close enough to the subscriber line of interest to couple the interfering signal into the line. Bit allocation and subchannel gains can be changed for communications in either direction. That is, it may be changed for upstream, downstream, or both communications. In effect, this allows
Ensure that the sub-channel capability matches the selected data rate and does not exceed the pre-specified bit error rate. When the jamming event ceases, the system is readjusted to its initial state of high rate.

Disturbing Events Of particular importance to the present invention are disturbing events resulting from the simultaneous transmission of data over the link and voice communication activity over the data link. These activities include activities such as the voice communication itself, or signaling associated with such communication, and the response to such activities, such as taking the phone off-hook or taking the phone on-hook. Including. Jamming events also include other disruptive jamming, such as interference from adjacent telephone lines caused by the presence of other DSL services, ISDN services, TI services, and the like. The interruption of the disturbing event itself may also include the disturbing event. For example, changing a voice communication device, such as a telephone, from an “on-hook” to an “off-hook” state is not compensated as described herein, or otherwise as previously done. P
Unless separated from the modem by the OTS splitter, it can severely disrupt communication with the modem. Therefore, this is a sabotage event that must be dealt with. However, returning such a device to the "on-hook" state is also a disturbing event that must be dealt with as it can significantly change the channel characteristics. The invention described herein efficiently addresses these and other disturbing events.

Channel Control Parameter Set In accordance with the present invention, bit allocation changes are made by switching between a stored parameter set containing one or more channel control parameters that define data communication by a modem over a subchannel. Achieved quickly and efficiently. The parameter set is preferably determined at modem initialization and stored in registers or other memory (eg, RAM or ROM) of the modem itself, but instead is external to the modem and communicates with the modem. However, it may be stored in a device such as a personal computer or on a disk drive.

According to one embodiment of the present invention, the channel control parameter set is stored at least in a primary channel control table to define a primary set of channel control parameters defining communication in the absence of voice communication activity or other interference. And one or more secondary sets of channel control parameters stored in the secondary channel control table and defining data communication in response to one or more jamming events. Hereinafter, when communicating under the control of the primary channel control table, the modem is described as being in the "primary" state. Hereinafter, when communicating under the control of a secondary channel control table, a modem is described as being in a "secondary" state. The modem is switched between a parameter set in a primary state and a parameter set in a secondary state in response to the occurrence and termination of a jamming event and is responsive to a change from one jamming event to another. And switched between the parameter sets of the secondary table. Since the parameter sets are pre-stored and therefore do not need to be exchanged in case of a jamming event, the switching takes place quickly, a jamming event is detected and typically at best a participant in a communication such as a second modem. It is essentially limited only by the speed with which the modem is signaled. This greatly reduces the disruption of communication that would otherwise be required by a complete re-initialization of the modem and typically associated switch channel control parameters over 6 to 10 seconds. ..

As mentioned previously, in DSL communications, information transmission typically occurs in both directions. That is, the upstream or ATU-C modem transmits downstream to the ATU-R modem via the first set of subchannels, and the downstream or ATU-R modem transmits via the second different set of subchannels. And transmits upstream to the ATU-C modem. Thus, the transmitters and receivers of each modem are used by those transmitters and receivers in transmitting data to and receiving data from the other modem forming a communication pair. Maintain the corresponding channel table. Certain parameters, such as time and frequency domain equalizer coefficients and echo canceller coefficients, are "local" to the receiver with which they are associated, and thus need only be maintained at that receiver. Other parameters, such as bit allocation and channel gain, are shared with the other modem ("modem pair") with which a given modem is communicating and are therefore stored on both modems. As such, during a given communication session, the transmitter of one modem uses the same set of shared parameter values as the receiver of the other modem, and vice versa.

Specifically, in DSL communication, the main parameter is the number of bits transmitted via various subchannels. This is known as the "bit allocation" of each subchannel and is a key element of the primary and secondary parameter sets. This is the channel SNR of the subchannel, the acceptable bit error rate,
And in a known manner for each sub-channel based on the noise margin. Another important factor is the gain of each sub-channel, so this factor is also preferably included in the primary and secondary parameter sets. Thus, each receiver stores a primary channel control table and a secondary channel control table, each table being used by that receiver and the transmitter of the other modem in communication with that receiver. Includes one or more parameter sets that define the subchannel bit allocation to be performed. And each transmitter also stores a primary channel control table and a secondary channel control table, each table being used by that transmitter for transmission to the other receiver and reception at that receiver. Specifies the sub-channel bit allocation and gains that will be applied. Part of each receiver's primary and secondary channel control tables that most closely matches the actual line they communicate with, which defines the parameters used in transmitting to a particular receiver. Preferably, as described herein,
Although determined by the modem in which the receiver is located (“local modem”), it will be understood from the detailed description herein that such a table may be determined in other ways.

Unless the communication over the subscriber line is compromised by a jamming event, the modem uses the primary channel control table to define the communication over the subchannels. However, when a jamming event occurs, the modem that detects the event (referred to herein as a "local modem", typically this is the subscriber, especially in the case of activation of the voice communication device by the subscriber. Modem ATU-R) informs the other modem that it needs to change to the secondary channel control table, and if there is more than one bit allocation set and / or gain set, it Identify the particular bit allocation set and / or gain set in the following table. The notification procedure is described in more detail below. Communication then continues according to the appropriate parameter set (ie, bit allocation, sub-channel gain, and possibly other parameters) from the secondary channel control table. This state is
It continues until a new jamming event is detected, and when a new jamming event is detected, the modem either returns to the primary channel control table (if the jamming is simply jamming or jamming the communication), or Return to a different parameter set secondary channel control table (if the jamming event is another jamming or occurrence of jamming that disrupts communication).

In addition to changing the bit allocation between sub-channels and changing the sub-channel gain, further changes may be made to communication parameters such as time domain equalizer coefficients, frequency domain equalizer coefficients, etc. Good. These parameters may also be stored in a channel control table used in controlling communication or in a separate table. In addition, power level changes (and corresponding changes in bit allocation and other communication parameters) for upstream and / or downstream communication may also be made and for use in controlling communication. At these power levels, a set of control parameters may also be defined. These changes are described in more detail below.

As currently contemplated, each modem on the subscriber side of the DSL line communicates with a corresponding dedicated modem on the central office side. Therefore, each central office modem (ATU-C)
Only the particular subscriber needs to store the primary and secondary tables. However, efficiency can be achieved whenever it is not necessary to service each subscriber at all times. Under these circumstances, the central office modem may be shared between two or more subscribers and switched between those subscribers as needed. In such cases, the ATU-C stores or has access to a set of channel control tables for each subscriber modem served by the ATU-C.

Table Initialization In the preferred embodiment of the present invention, the primary and secondary channel control tables are determined in an initial "training" session ("modem initialization"). In this session, known data is transmitted by one modem, measured on reception by the other modem, and a table is calculated based on these measurements. Typically, training sessions occur when a modem is first installed in a subscriber's home or central office, and thus the procedure "specializes" the modem into the environment in which it operates. This environment includes, in addition to the subject data modem, one such as a telephone handset, a facsimile machine, and other such devices that communicate via voice frequency subchannels, typically in the range of 0-4 kHz. The above voice communication device is included. A primary channel control table comprising a parameter set including at least a set of sub-channel bit allocations, preferably also including a sub-channel gain, is calculated with each device inactive. A secondary channel control table containing one or more bit communication parameter sets (bit allocation, gain, etc.) is provided with each voice communication device activated separately and / or with a group of devices activated simultaneously. Calculated by The table so determined is then stored at the receiver of one modem and further transmitted to the transmitter of the other modem and stored at that transmitter for use by both modems in subsequent communications. To be done.

Another approach is to determine the secondary channel control table (including one or more parameter sets containing the table) by calculation from the primary channel control table. This is most simply a percentage of the corresponding primary parameter, eg, one or more of the parameters (eg, a bit allocation parameter that defines the number of bits used for communication over each subchannel). , The SNR of each sub-channel, or as a constant or varying percentage over the sub-channels.
Achieved in response to a decision according to a constant or varying percentage over sub-channels, or according to a decision according to a bit error rate different from the bit error rate provided in the primary channel control table. Or achieved by other techniques.

As a specific example, a number of different bit allocation sets in the secondary channel control table are:
It may be determined as a different percentage of the corresponding bit allocation set in the primary channel control table (constant or varying across subchannels). Each secondary bit allocation set corresponds to the effects typically caused by a particular device or class of devices, such as telephone handsets, facsimile machines, etc., as determined by repeated measurements on such devices. Thus, it may be considered to represent the expected impact of the device over a range of communication conditions, such as using a particular type of subscriber line wiring, being in a given range from the central office, and so on. Then, the sub-channel gain may also be adjusted based on the re-determined bit allocation. The bit allocation and subchannel gains thus determined form a new secondary parameter set. These new quadratic parameter sets can be used in response to the detection of the jamming event that these new quadratic parameter sets characterize, as well as the determination of the quadratic bit allocation and gain based on the measurement of the actual jamming being compensated. Is a substitute for.

Alternatively, the secondary channel control table is
It can be determined by adding a power margin to the calculation for each entry in the primary table that is large enough to accommodate interference from voice communication device activation or from other interference. This has the effect of reducing the constellation size of table entries. As well as the percentage factor when that approach is used,
This margin may be uniform across table entries or may vary across table entries. Multiple secondary bit allocation sets, each based on a different power margin, may be defined by this approach.

One example of the use of varying margin is crosstalk (capacitively coupled noise for nearby xDSL users, where "x" is ADS).
L, HDSL, etc., which indicates the possible types of DSL). This crosstalk is generally more predictable than the signaling events associated with voice communications.
The crosstalk spectrum of the xDSL source is well characterized. For example, American Natural
TI. published by the Nal Standards Institute. See 413 ADSL Standard. From the primary channel control table associated with one complete initialization, a secondary table can be calculated containing a family of bit allocation sets, each bit allocation set corresponding to a different crosstalk level. As the number of xDSL systems (and thus crosstalk levels) changes, the ADSL link can quickly switch to one of these automatically generated sets.

The secondary channel control table of the present invention also makes measurements on the transmission information of each superframe during, for example, data communication, and monitors these measurements when channel performance changes significantly. , Different bit allocation sets and possibly different gain sets should be used to dynamically adapt. Applicants have discovered that SNR provides an easily measurable and reliable indicator of required bit allocation and gain.

Specifically, Applicants have determined that the measurement of SNR levels across multiple subchannels during a given communication condition or state is such that the set or gain of bit allocations used in subsequent communications during that state. It has been discovered that it provides a "fingerprint" that can be reliably used to quickly select a parameter set, such as These measurements are, for example, the sync that occurs in each superframe.
A frame, and more generally, a sync frame, which occurs during transmission of a reference frame. SNR
However, if during communication more than a specified amount is changed, the modem in which the measurement is performed searches the stored parameter set for the set whose SNR on the corresponding subchannel is closest to the measured SNR, and then Select a set,
Used in subsequent communications. If no parameter set is found within the defined range, the system may be switched to a default state or correspond to a defined SNR pattern over some or all of the sub-channels that should be used. Complete re-initialization may be required. SNR measurements can also be made on the data that carries the signal itself. That is, the decision-directed S
NR measurement.

Instead of using multiple secondary subchannel control parameter sets as described above, a simplified approach is based on a composite of individual device bit allocations or other characteristics to create a single binary channel. The next set may be configured and used. In one embodiment, the composite is formed by selecting for each sub-channel the least significant bit allocation exhibited by any device in that sub-channel or any other most significant characteristic of interference. , Thereby forming a single "worst case" set that can be used when any device is activated, regardless of the particular device or interference actually present. Alternatively, the composite is the actual or calculated capacity of the line when all devices are actuated at the same time or in theory, or when all disturbances are present, or both. May be determined as Bit allocation sets may also be determined for combinations of such devices and subsets of jammers. A similar approach may be used to address the situation where several devices are activated at the same time, and the effects of other disturbances such as crosstalk may also be incorporated into the composite set.

The particular parameter set of the secondary channel control table may be set for the duration of the session in which the voice device is active, or for example another voice device is activated or some other interruption occurs. , Remains in use until another state change occurs. When this happens, the local modem resumes its identification procedure, allowing the determination of the appropriate parameter set for the new condition. At the end of the session where the voice device is active, the device returns to the inactive (ie, "on-hook") state, and the system returns to the initial ("on-hook") state. In this state, the primary channel control table is
Once again used for communication between the central office and the subscribers.

The switching of subchannel parameter sets according to the invention is very fast. This switching is accomplished at intervals as short as a few frames, thus avoiding the long (eg, several seconds) delays that would otherwise be associated with determining, communicating and switching the newly determined set.
Furthermore, this switching avoids the communication of new parameter sets when the communication is compromised and the error rate is high. Therefore, this switching minimizes the disruption of the communication process caused by disturbing events.

Detecting Jamming Events During subsequent data communication, identification of activated devices is accomplished in one of a number of ways. In one embodiment of the invention, when a device is activated, a specific activation signal is sent from the device to a modem on the same side of the subscriber line as the device (referred to herein as a "local modem"). ) Is transmitted to. This signal may be transmitted over a communication line to which the device and the local modem are connected, or it may be sent via a dedicated connection between the device and the local modem.

In the preferred embodiment of the invention, the local modem monitors the subscriber line to which the local modem and the device are connected, and detects changes in line characteristics when the device is activated. For example, the signal-to-noise ratio (SNR) of various subchannels can be quickly measured,
It can then be used to identify the particular device to be activated. During multiple sets of initializations corresponding to multiple communication conditions caused by a device or other interference, the SNR measurements for each sub-channel may be tracked under conditions that are tracked (ie, no device is activated, device Are separately activated, two or more devices are simultaneously activated, adjacent channel interference, etc.), and these measurements are determined by the singular or It is stored with the identification of multiple specific parameter sets. When a device is activated, SNR measurements can be used to quickly identify the specific device or devices being activated, and then the local modem can switch to the appropriate secondary table.

Interfering events can also be detected in accordance with the present invention by monitoring selected transmission characteristics dependent on these events. These jamming events include, in addition to any specific SNRs associated with these jamming events, errors in the cyclic redundancy code (CRC) associated with the transmission and changes in the error rate of this code, amplitudes of pilot tones on subchannels. ,
It may include measurements such as changes in frequency or phase, or other such indications. Forward Error Correction Codes (FEC) are typically used in ADSL transceivers and errors in this code such as how many errors have occurred, how many errors have been corrected, and how many errors have not been corrected. Changes in rate characteristics can be particularly useful in detecting jamming events.

In monitoring these characteristics, Applicants have found that lightning or other such burst noise disturbances,
Distinguish between changes caused by momentary or transient events such as, and changes associated with disturbing events. The latter lasts for a considerable interval (eg, on the order of seconds or more). Specifically, in embodiments that monitor CRC error or error rate in accordance with the present invention, switching from one parameter set to another may occur if the error spans multiple frames or if the error rate is at a certain time. Is provided during a change in the specified amount by more than a specified minimum value during. For example, when an off-hook event occurs, which is a significant form of disruption to data communication over the subscriber loop, the number of CRC errors suddenly increases and remains at the increased level until addressed. This is distinguished from the occurrence of transient disturbances, such as lightning strikes, which cause a momentary increase in CRC errors that do not persist until the system loses synchronization.

Therefore, according to the invention, the detection of an initial change in the CRC error rate over a defined threshold over a large number of frames is one implementation of the detection of a jamming event which results in the switching of parameter sets. Here is an example.
A similar procedure may be undertaken in response to measuring the subchannel signal-to-noise ratio to detect jamming events based on this property. A determination as to whether a jamming event has occurred is made by measuring on a single sub-channel, measuring on multiple sub-channels (e.g. if more than a specified number of sub-channels detects a jamming event, the The decision to switch is made), and so on.

Another technique for detecting jamming events in accordance with the present invention is the use of monitor signals such as pilot tones whose amplitude, frequency, phase or other characteristics are monitored during data transmission. Sudden changes from one or more frames of monitored characteristics to another frame,
Then, in subsequent frames, if smaller changes or no changes occur, this indicates a jamming event to which the modem should respond. The monitor signal can be a dedicated signal carried by one of the sub-channels, a signal carried on a separate control sub-channel, the jamming event itself (eg, the presence of ringing tones, dial tones, or other common telephone signals). ), Or other signals.

Communicating the Occurrence of a Jamming Event Once a jamming event has been detected and the appropriate parameter set corresponding to that event has been identified, the identification is communicated to the remote modem by a select signal, and the remote modem also Allows you to switch to the corresponding parameter set in the next table. The selection signal may be in the form of a message transmitted over one or more sub-channels or using a predetermined protocol for the embedded operating channel, or the selection signal may be a specific signal. It may include one or more tones that identify the parameter set. ADSL systems use a "guard band" of several subchannels between the set of subchannels used for upstream and downstream transmission. This guard band may be used to carry one or more selected tones. If there is only one parameter set specified, the select signal is a simple flag sent to the remote modem to select that set (has only two states: on / off, on / off). , Etc.) may be included.

In another embodiment of the invention, ATU-R and ATU, commonly provided in DSL systems.
-The frame counter in the C modem is used. Upon detecting a jamming event, the ATU-R modem notifies the ATU-C modem of the event and changes the parameter set,
Alternatively, it identifies the frame in which the power level change, or any change in other parameters associated with it, occurs. This identification may be direct (ie the notification specifies a particular frame number in which changes to the secondary table are made) or indirect (ie the notification receives One of a predetermined number of frames, eg "0"
Or a change to the secondary table is made at the next frame number ending in "00", etc., or at the nth frame after receiving the notification, where n is some number greater than 0). .. When the specified frame is reached, both modems (ie ATU-R and ATU-
C) switches to the new bit allocation set, power level, and other specified parameters.

Alternatively, upon detecting a jamming event, the modem performs "fast retraining" to characterize the communication under the new operating conditions and to determine the power and / or bit allocation set used for the communication. . Fast retraining
Only a limited subset of the full initialization procedure is performed, eg bit allocation and sub-channel gain determination. The retraining modem (typically the one that initiates retraining and is local to the jamming) then compares the newly determined parameter set to the previously stored set. If the newly determined set is the same as the previously stored set, one modem communicates a message, flag, or tone to the other modem,
Specifies which of the stored secondary allocation sets will be used. Otherwise, the newly determined set is used for communication. In the latter case, this set must be communicated to the other modem in the communication pair, and communication can be interrupted while this is happening. Nevertheless, when an event that requires a change in a parameter set ceases, the system may simply revert to the primary parameter set without having to retransmit the set and thus without further interruption of communication. With proper care in initialization, in most cases sufficient parameter set arrays can be initially defined and replaced to avoid the need for subsequent re-initialization in response to most disturbances.

In addition to changing one or more parameter sets in the modem in response to a power level change jamming event, according to a preferred embodiment of the present invention, it is preferable to further enhance reliable communication. Changes the communication power level in the upstream and / or the downstream direction. Typically, this change is to reduce the power level in the upstream direction to minimize interference to voice communications and to reduce echo to downstream signals. In the book, it is explained as such. However, interference from adjacent data services
Increasing power levels may also be required, such as when higher power levels are required to maintain a desired data rate or bit error level, and such changes may be reduced by the present invention. It should be understood that the same is achieved. Further, changes in the downstream power level may be required if the line conditions change such that excess power would be supplied to the downstream channel from the upstream modem without changing the downstream power level.

Theoretically, and in a perfectly linear system, upstream communication activity should not affect concurrent voice communication. This is because the two activities occur in separate, non-overlapping frequency bands. However, the telephone system is not really a linear system, and the non-linearity of the system allows and can actually inject the image signal from the upstream subchannel to the audio subchannel and possibly also the downstream subchannel. (Ie, echo), which causes detectable interference. In accordance with another aspect of the invention, this effect can be seen when the state dictates the upstream power level, for example when the voice communication device is off-hook and a leak from the data communication being conducted interferes with the voice communication. By reducing the (power level transmitted by the subscriber or downstream modem to the central office or upstream modem) by a given amount or factor, it is reduced to a level below the undesired level.

The power reduction amount may be set in advance. For example, Applicants have found that it is most common to reduce this power by 9db (relative to the power used in ADSL applications that typically use splitters to separate the data and POTS signals). Found to be sufficient in important cases. Under these circumstances, the system will always operate at two other power levels. Alternatively, the downstream modem may select one of several different power levels to be used based on the then prevailing communication conditions resulting from the jamming event. For example, a downstream modem may send a test signal to one or more upstream subchannels, and may leak (ie, echo) this signal to one or more downstream subchannels, eg, as determined by the SNR on the subchannels. ) Can be activated. The power level transmitted by the downstream modem upstream may then be adjusted accordingly to minimize the effects of echo. Generally, the downstream transmission power is determined by ATU-R. Because AT
This is because UR is the closest to the cause of the disturbing event. In this case, the ATU-R uses a message, flag, or tone to inform the ATU-C of the desired power level used for transmission. In either case, at the end of the session, the power level returns to the power level used in the "on hook" state.

In selecting the desired power level, the transmitting modem includes in the receiving modem of the communication pair the desired modification (where appropriate, designation of a particular power level from among several power levels). ), And then implement the change, including switching to a new parameter set associated with that power level. In another embodiment of the invention, the receiving modem detects a power level change at the transmitting modem and switches to a parameter set associated with that power level. Thereafter, upstream communications (ie, ATU-R to ATU-C) occur at the new power level until the jamming event (eg, off-hook condition, etc.) is over.

Although much of the above has been described with respect to changing the power level of upstream communications from a subscriber modem to a central office modem, it should be noted that changing the power level in the opposite direction may sometimes be required. This may be the case, for example, with short subscriber loops (eg less than a mile). In this case, the central station may initially transmit at excessive power levels as a result of the reduced line loss that results from being closer to the central station. In such a case, the central office or ATU-C modem may have previously
The R-modem plays the role it played, and vice versa. And the modification of the power level and other parameters in the downstream communication can be done as described above. Moreover, power changes are most commonly expected to be changes that reduce the power level used to communicate, although in some cases power changes increase power levels. It should also be understood that there are good things. This occurs, for example, when crosstalk from adjacent services requires an increase in the power level of the service of interest to compensate for the crosstalk.

Modification of Other Parameters Another important modification made in response to the detection of a jamming event is:
A frequency domain equalizer (“F
DQ ”). These equalizers compensate for various distortions (e.g., amplifier loss, phase delay, etc.) experienced by the data during transmission through the subchannels. Typically, these equalizers include finite impulse response filters with complex coefficients. These factors are set during the "initialization" or "training" phase of modem setup. These coefficients may then be adjusted based on the reference (known) data of the reference frame or sync frame transmitted over the communication subchannel. According to the invention, these filters are adjusted in response to the transmitted reference data when a disturbing event is detected. The coefficient updates may be done for all subchannels or selectively for those subchannels where changes in error rate, signal-to-noise ratio, or other error signature indicate jamming events. Good.

According to one embodiment of the present invention, both when there is no jamming event or jamming ("first order FDQ coefficient") and when there is such an event or jamming ("second order FDQ coefficient"). The coefficients of the frequency domain equalizer for the communication of are calculated and stored during initialization or training. Thereafter, as in the case of the channel control table, these coefficients are switched in response to a disturbing event and then reverted to their initial state when such an event stops.

According to another embodiment of the invention, the FDQ coefficients are recalculated in response to the detection of the jamming event, and then
It is used for the rest of the communication session instead of the previously stored secondary FDQ table. Recalculation is accomplished in a short "retraining" session in which known reference data is transmitted between ATU-R and ATU-C. The received data is compared to the known data and the new FDQ coefficients are determined accordingly. In addition to frequency domain equalizer coefficients, time domain equalizer coefficients and echo cancellation coefficients can also be determined and stored. Such coefficients are local to the particular receiver and therefore need not be communicated to the other modem of the communication pair. Therefore, any such retraining is very fast and limits any resulting communication disruption.

Excessive Interference In some cases, a particular device may cause interference to communications such that compensation of that device by the methods described herein is not useful. This can occur, for example, in old phones or in particularly complex home wiring. In such cases, it is desirable to insert a simple in-line filter between the device and the subscriber line to minimize the interference caused by such device. The filter may include, for example, a simple low pass filter with a volume of 1 cubic inch or less, and a pair of standard connectors such as the RJ11 connector that connects one side of the filter to the device and the other side to the subscriber line. . Unlike POTS splitters, such a connector does not require a trained technician to install the connector, thus impeding the acceptance of ADSL modems as described herein is an obstacle and cost. Or give no other. Such a device can be detected by measuring the non-linear distortion of the device when the device is activated. This is done by monitoring the echo on the line caused by the device.

Reduced Rate Communication A further improvement in the operation of the modem of the present invention is to limit the bandwidth of downstream transmissions to a subset of the bandwidth normally provided by ADSL communications. This reduces the processing requirements for both local (ie central office) and remote (subscriber home) modems, thereby allowing subscriber home modems to be priced more acceptable for customer non-business use. Make it easy to provide. Moreover, this further minimizes the interference between data transmission and voice communication. For example, limiting the number of sub-channels used by the modem to 128 instead of 256 reduces the downstream bandwidth from 1.1 MHz to about 552 kHz. If the modem is typically used with a modem that provides a higher number of subchannels for such communication, the bit allocation and gain of more than 128 subchannels is preferably zero. That is, it is set to zero.

The present invention is preferably operable not only with modems that do not have the capabilities described herein, but of course is capable of operating with modems having such capabilities. Thus, the modem of the present invention identifies the capabilities of a modem prior to exchanging data with another modem, preferably during initialization. According to a preferred embodiment of the invention, this is preferably done by signaling between the modems participating in the communication. This signaling identifies the type of modem with which it is communicating and the characteristics of the modem that are important to the communication session. For example, one form of ADSL transceiver uses a reduced number of subchannels (typically 32 subchannels upstream and 128 subchannels downstream), and a lower bandwidth. Provide communication. Then a full AD encountering a reduced rate modem
A modem with SL capability can adjust its transmit and receive parameters to match its reduced rate modem. This can be done, for example, by transmitting tones reserved for such purposes from one modem to another.

In particular, according to the present invention, when communication is initiated between the central office modem and the subscriber's home modem, these modems will be able to call themselves "complete". Rate "(ie communicating over 256 sub-channels) or" reduced rate "(eg communicating over a somewhat smaller number of channels, eg 128 sub-channels) ). The communication is performed by a flag (2 states, for example, "on / off", "present / off"
None ”) may be done via one or more tones, messages (n states, n> 2), or other forms of communication, and initiated at either end of the communication subchannel. Good. That is, it may start at the end of the central office or at the end of the customer's house.

(Detailed Description of Exemplary Embodiments) The following description of the invention refers to the accompanying drawings.

FIG. 1 shows an ADSL communication system of the type previously used which incorporates a "splitter" to separate the voice and data communications transmitted over the telephone line. As shown in FIG. 1, a telephone central office (“CO”) 10 is connected by a subscriber line or loop 14 to a remote subscriber 12 (“CP: customer premises”). Subscriber line 14 typically includes twisted copper wire pairs. It has been the traditional medium for carrying voice communications between telephone subscribers or customers and central offices. Designed to carry voice communications with a bandwidth of about 4 kHz (kilohertz), its application has been greatly expanded by DSL technology.

The central office is then connected to a digital data network ("DDN") 16 for transmitting and receiving digital data, and a public switched telephone network ("PSTN" for transmitting and receiving voice and other low frequency communications. ]) 18 is connected. The digital data network is connected to the central office via a digital subscriber line access multiplexer (“DSLAM”) 20 and the switched telephone network is connected to the central office via a local switch bank 22. The DSLAM 20 (or its equivalent, such as a data-enabled switch line card) connects to a POTS "splitter" 24 via an ADSL transceiver unit-central office ("ATU-C") 26. The local switch 20 also
Connect to splitter.

Splitter 24 separates the data and voice (“POTS”) signals received from line 14.
At the subscriber end of line 14, splitter 30 performs the same function. Specifically, the splitter 30 connects the line 14 to a suitable device, such as a telephone handset 31, 32.
Send OTS signal and AD digital data signal
SL Transceiver Unit-Subscriber ("ATU-R")
Sent to a personal computer (“PC”) 36
Etc. to the data utilization device. Transceiver 34
May preferably be incorporated as a card on the PC itself. Similarly, transceiver 26 is typically implemented as a line card for multiplexer 20.

In this approach, a communication channel of given bandwidth is divided into a number of subchannels. Each subchannel is a portion of the subchannel bandwidth. 1
The data transmitted from one transceiver to another is modulated according to the information carrying capacity of a particular subchannel and provided to each subchannel. Due to various signal-to-noise (“SNR”) characteristics of subchannels, the amount of data loaded on a subchannel may vary from subchannel to subchannel. Accordingly, a "bit allocation table" (shown as table 40 at transceiver 26) for each transceiver is provided to define the number of bits that each transceiver transmits on each subchannel to the receiver to which it is connected. Table 4 with transceiver 34
2) is maintained. These tables are created during the initialization process. The initialization process involves each transceiver transmitting a test signal to the other transceiver, and then measuring the signal received by each transceiver and transmitting it from one transceiver to the other on a particular line. Determine the maximum number of bits you can have. The bit allocation table determined by a particular transceiver is then transmitted over the digital subscriber line 14 to the other transceiver, which in turn receives that particular transceiver or any similar device connected to the line 14. Use this when transmitting data to the transceiver. Needless to say, the transmission must take place when the line is undisturbed which could interfere with the communication. This is a serious limitation and limits the use of this approach.

Referring now to FIG. 2, the bit allocation table 42 as used in the customer premises equipment.
Are shown in more detail. The table 40 used at the central office is essentially the same in structure and operation,
No further explanation is given. The table 42 has two parts. The first portion 42a defines certain communication parameters, such as bit allocation capabilities, that characterize each sub-channel, and sub-channel gain parameters, and the transmitter portion of the transceiver 34 is the transceiver 3
The first part 42a is used in transmitting signals to the other transceiver (26) with which 4 is in communication. The second part 42b defines the parameters used by the receiver part of the transceiver 34 in receiving the signal transmitted from the other transceiver. Communication occurs over multiple sub-channels, shown here as sub-channels “9”, “10”, etc. in the transmitter section and sub-channel “40” in the receiver section, for illustrative purposes only. , "41" and so on. In a full rate ADSL system, there are up to 256 such sub-channels, the bandwidth of each sub-channel is 4.1 kHz. For example, in one embodiment of the present invention, upstream communications (i.e., from customer premises to central office) are subchannels 8-29.
And downstream communications (from the central office to the customer's premises) occur on subchannels 32-255. Sub channel 30
And 31 form a guard band between upstream and downstream communications, which may be used for signaling as described below.

For each sub-channel (“SC”) 50, field 52 contains a communication pair or modem according to the prevailing conditions of the sub-channel, eg, measured signal-to-noise ratio (SNR), desired error rate, etc. It defines the number of bits ("B") transmitted by the pair of transmitters via the subchannel and received by the pair of receivers. Column 54 defines the corresponding gain ("G") of the sub-channel. First portion 42a of table
Identifies the bit allocations and gains that the transceiver 34 uses in transmitting "upstream" to the transceiver 26, and the second portion 42b defines the bit allocations that the transceiver 34 uses in receiving the transmission from the transceiver 26. And specify the gain. Transceiver 2
6 is a corresponding table 40 which is a mirror image of the table 42.
Have. That is, the bit allocation specified for transmission by transceiver 34 is the same as the bit allocation specified for reception by transceiver 26,
The same applies to reception by the transceiver 34 and transmission by the transceiver 26. The table may also typically include fields that identify the gain 54 associated with a particular subchannel.

As mentioned above, the splitters 24, 30 are
The data and voice communications applied to the splitters 24, 30 are combined and transmitted, and once received the data and voice communications are separated from each other. This is accomplished with high pass and low pass filters that separate low frequency voice communications from high frequency data. But,
The need to use such splitters seriously impedes widespread adoption of DSL technology by customers. Specifically, a trained technician must go to the house to install the splitter in the subscriber's house. This can be quite expensive and discourages many, if not most, customers from using the technology. Incorporating a splitter into the communication device itself is also not a viable option. Because this not only increases the cost of such a device,
It requires either the purchase of all new equipment or the modification of older equipment. This modification also requires skilled help to achieve it. The present invention eliminates at least the customer premises splitter, thereby allowing end users to adopt and use DSL modems without the intervention of a skilled technician. But,
This requires significant changes in the structure and operation of the DSL transceiver or modem and the present invention addresses these changes.

Specifically, FIG. 3 shows a DSL according to the present invention.
1 shows a transmission system. In this system, a composite voice-data signal transmitted from a central office to a subscriber's house does not require a subscriber's house splitter to intervene with the subscriber's voice equipment 31, 32 and a data transceiver or modem 3.
4'and sent to both. In FIG. 3, the components that are unchanged from FIG. 1 have the same numbers, and the modified components are indicated by the prime numbers in the upper right. Instead of the single table 30 of the transceiver 26 of FIG. 1, the transceiver 26 ′ of FIG. 3 includes a primary channel control table 41 and a secondary channel control table 43. Similarly, the transceiver 34 ′ of FIG. 3 has a primary channel control table 45 and a secondary channel control table 47.
Including and Note also that in FIG. 3, the subscriber side splitter is eliminated. The reason why this can be done with the present invention will now be explained in detail. Note also that the central office splitter 20 of FIG. 1 is retained in the configuration of FIG. This is optional and not mandatory. Retaining the central office splitter can provide some performance gains at a small cost, since generally only one installation at the central office itself is available to the technician at any given time. Otherwise, the central office may omit the splitter.

Referring now to FIG. 4, the transceiver or modem 34 'is shown in more detail. Modem 26 '
Are essentially the same for the purposes of the present invention and will not be described separately. As shown, the modem 34 'includes a transmitter module 50, a receiver module 52, a control module 54, a primary channel control table 45, and a secondary channel control table 47. The primary channel control table is more fully shown in Figure 5A and the secondary channel control table is more fully shown in Figure 5B.

In FIG. 5A, the primary channel control table 4
5 is used to receive a communication from the remote transmitter over the DSL line and a transmitter section 45a that stores a primary set of channel control parameters used when transmitting to the remote receiver over the DSL line. A receiver section 45b for storing a primary set of channel control parameters,
Have. The subchannel to which the parameter applies is shown in column 45c. The channel control parameters of the transmitter section 45a include at least the identification of the bit allocation ("B") 45d used on each sub-channel during transmission, and preferably also the identification of the gain ("G") 45e. Including. The receiver section also includes bit allocation and gain identification, preferably a frequency domain equalizer coefficient ("FDQ") 45f, a time domain equalizer coefficient ("TDEQ") 45g, and an echo canceller, among others. It also includes the identification of the coefficient (“FEC”) 45h.

Parameters such as bit allocation, gain, frequency domain coefficient, time domain coefficient, etc. collectively form a parameter set, and each element of this parameter set is also a set. For example, a bit allocation set that defines the allocation of bits to each subchannel, a gain setting set that defines the gain of the subchannel, and the like. According to a preferred embodiment of the invention, the primary channel control table is a single parameter set having at least one element, namely a bit allocation set, preferably also a gain allocation set. Stores the set. This set of parameters defines the default communication conditions under which the system will return in the absence of disturbing events. However, the secondary channel control table has at least two parameter sets, typically more than two, for controlling the transmissions and receptions performed by the respective modems over the subscriber line. . These sets define communication under various jamming events that change the default conditions.

Specifically, in FIG. 5B, the secondary channel control table 47 includes a plurality of parameter sets 47a and 4a.
7b, 47c, etc. are included. For illustration purposes, only three of these parameter sets are shown. Each parameter set includes a transmitter 47d and a receiver 47e. In each part, one or more parameters are specified. For example, the transmitter 47d has bit allocation 47f and gain 47g, and the receiver 47e has frequency domain coefficient 47h and time domain coefficient 47g.
i and the echo cancellation coefficient 47j. The actual values of the coefficients are typically complex numbers, so these values are simply represented in the channel control tables of FIGS. 5A and 5B, for example by letters such as "a", "b". . The parameter sets 47b and 47c and the remaining parameter sets are similarly configured. As with the primary channel control table, each parameter (eg, bit allocation) itself is a set of elements that define the communication conditions over at least some of the subchannels, and these elements apply to these subchannels. And helps characterize these subchannels.

The primary channel control table containing the bit allocation parameter set is generated in the usual way, ie during initialization (typically prior to the transmission of "working data" rather than test data). , Known data is sent to and received from a remote modem with which the target modem is communicating under conditions that should include the modem's default conditions. Typically, this is a condition where all interfering devices are deactivated, so that the highest data rates can be achieved, but the actual conditions are user selected. The data received at each modem is collated with the data known to be transmitted, and the primary channel control parameters such as bit allocation, subchannel gain, etc. are calculated accordingly. This table is then used as long as the system remains undisturbed by jamming events that disrupt communication over the line.

The secondary channel control table can be determined during initialization in the same manner as the primary table, but this determination is to redetermine the channel control parameters needed for communication under the new conditions. , Can be performed by activating a device that can cause a jamming event. These devices may be activated one at a time, a secondary parameter control set may be determined for each device and stored in the secondary channel control table, or they may be a group of two or more devices. Operated in units and the parameter set determined accordingly, or
Various combinations of single and group operations may be performed and the corresponding parameter set determined. The secondary parameter set may likewise be determined from actual measurements with interference sources such as xDSL transmission on a common binder with the relevant modem, and the resulting set may be stored in a secondary table. .

Other secondary table determination methods may be used. For example, one or more secondary parameter sets may be obtained from the primary table. Therefore, the bit allocation for each sub-channel of the secondary table may be considered as a percentage of the bit allocation for each sub-channel defined in the primary table, which may be a constant or a varying percentage across the sub-channels. Alternatively, the bit allocation may be calculated from the same data as in the primary table, but with a larger margin. Percentage of the signal-to-noise ratio used in calculating the primary table, with a constant or varying percentage over the subchannels, or considering a bit error rate different from that considered in the primary table Or calculated by other techniques, including those previously described. The primary and secondary parts may be recomputed or refined during the communication session and then stored for later use. The calculation or recalculation may be a single event,
It may occur repeatedly, such as occurring periodically throughout the communication session.

Moreover, the use of multiple parameter sets in the secondary channel control table generally provides the best match with the actual channel conditions, and thus closer to the optimal communication conditions, but with a single composite parameter. A simplified second table containing the set may be used. Thus, FIG. 5C illustrates multiple bit allocation sets 49a-49d for subchannel 49e. This may represent a corresponding number of different communication devices or conditions associated with communicating over these sub-channels. The single composite parameter set 49f may be formed as a function of parameter sets 49a-49d, for example, by selecting the smallest bit allocation for each sub-channel in sets 49a-49d for each sub-channel 49e. . Such sets represent "worst case" conditions for activation of any of the devices associated with sets 49a-49d. Other worst case parameter sets may be created, for example for selected device groups, in case several devices or disturbances are operating at the same time.

If there is no disturbing event, the transceiver 2
6 ′ and 34 ′ are primary channel control tables 41 and 45.
Use to communicate. However, in response to the detection of the jamming event, the transceivers 26 ', 34' switch to one of the parameter sets of the secondary channel control tables 43, 47 and continue to communicate under the conditions specified by the particular parameter table. . These conditions may specify a reduced bit rate while maintaining the same bit error rate as that provided in the primary channel control table, or a higher error rate but the same bit rate. May be specified, or the reduced bit rate may be specified with a correspondingly reduced power level or margin, or other conditions determined by a particular channel control table. Good.
When the jamming condition that caused the switch is over, the transceivers 26 ', 34' revert to using the primary tables 41, 45.

Typically, the primary table provides communication over or near the capacity of the communication channel over line 14. The secondary table provides communication over the channel at a reduced rate. Switching between the primary table and the secondary table according to the present invention (ie switching from the primary parameter set to the secondary parameter set) is fast. This switching can be accomplished at intervals as short as a few frames (currently A
In a DSL system, each frame is about 250 microseconds), and thus would otherwise be needed for decision, communication over the subscriber line, and switching of the newly determined bit allocation table. Avoid delays (eg, on the order of seconds). Furthermore, this avoids the communication of such tables over the subscriber line when the communication is compromised and thus the error rate is high. Therefore, the use of pre-stored parameter sets according to the present invention minimizes the disruption of the communication process caused by disturbing events.

The channel control table is stored in storage or memory for fast access and retrieval. Preferably, the storage device is a random access memory (“RAM”) built into the modem itself, but in a computer, such as in stand-alone memory, a personal computer (“PC”), in a disk drive, or other element. Also included is such memory located within other components accessible to the modem, such as within. In addition, the storage device may include other forms of memory such as read only memory (“ROM”).

In addition to accessing the channel control tables 45 and 47, the control module 54 of FIG.
Preferably, if the secondary control table is calculated based on the primary channel control table, it controls the formation of the secondary control table. Moreover, as in the most general case, the module 54 has a SNR on the subscriber line 14.
, And if the primary and secondary channel control parameter sets are based on measurements of the actual conditions of the line,
Calculate these sets. For this purpose, the control module is advantageously implemented as a dedicated digital computer or "DSP" chip specialized for the functions described herein. Alternatively, the control module may, of course, be implemented as a general purpose computer or in other ways, as will be appreciated by those skilled in the art.

According to the present invention, a jamming event on the subscriber line is distinguished by its consequences from a transient event such as a lightning impulse. Specifically, signaling events such as off-hook signals or on-hook signals typically
Cause enough confusion to prevent further communication without re-initialization. This event is accompanied by error code indications that persist throughout this disruption, changes in the amplitude and phase of the physical signals or pilot tones carrying the data, the application of significant voltage to the line, and other symptoms. . To detect the event, the subscriber line is monitored for the occurrence of one or more of these characteristics.

FIG. 6 illustrates one manner of detecting jamming events in accordance with the present invention. A detector 70, preferably included in control module 54, receives the signal from line 14 and
The error code associated with the signal (eg, CRC error or FEC error count) is then monitored (step 72) to see if any indication of an error has occurred. If no errors are detected (step 74), the detector remains in monitoring mode with no further activity. If the error code indicates an error, the counter is incremented (step 76) and the count is then compared to a predetermined threshold (step 78). If the count does not exceed the threshold value (step 78, ">N").
? )), The system remains in monitoring mode and continues to accumulate any detected errors. If the count exceeds the threshold (step 78, Y), the detector issues a "jamming event" detection signal (step 80). This detection signal causes the transceiver in which the detector 70 is located to initiate the process of switching to the appropriate parameter set in the secondary table. When this happens, the count is reset (line 81).

Instead of monitoring the error code for characteristic behavior (ie, errors that are repeated over successive frames), according to the present invention, the physical signal carrying the data over the subchannel is The amplitude and phase, or the amplitude and phase of pilot tones transmitted between modems may be monitored. When a jamming event occurs, the amplitude and phase of the physical signal undergoes significant changes. That is, the amplitude suddenly decreases and the phase suddenly changes to a new value. After that, the amplitude and phase maintain approximately their new values during successive frames. This behavior can be monitored as shown in FIG. Here, monitor 100 monitors, for example, the amplitude of a data signal or pilot tone on line 14, and upon detecting a change in amplitude greater than a predetermined threshold, flip-flop 102 is placed in the "active" state ( "Q"). Flip-flop 102 enables counter 104, which is connected to receive count pulses from frame counter 106 (input "E") whenever a new frame is sent to or received by the modem. These count pulses are also applied to the threshold counter 108. The threshold counter 108 accumulates the count provided to the threshold counter 108 until the count reaches a defined count and then provides the resulting count to the comparator 110. In the comparator 110, this count is compared with the count of the counter 104. If the contents of the counters 104 and 108 are equal, the comparator 1
10 provides the output ("Y"). This output causes the transceiver to begin the process of switching to the appropriate table. This also resets counters 104, 108 and flip-flop 102. Counter 10
4, 108 and flip-flop 102 are also reset (input "R") if the counts of counters 104 and 108 do not match ("N" output of comparator 110).

A similar procedure can be used to generate the table switching signal based on the monitoring of the phase change of the data signal or pilot tone as described above. Further, while the operation of the event detector of FIG. 8 is largely described in terms of hardware, the operation of the event detector is software-based, as is most of the elements described herein. However, it is understood that it can be easily realized by a combination of hardware and software.

Yet another approach to detecting jamming events is to directly monitor them. For example, for off-hook or on-hook signals, d of 48 volts
A c-step voltage is applied to the subscriber line. This signal can be easily detected directly by simply monitoring the line for a step voltage of this magnitude, and then generating a table switch signal in response to detecting that voltage. It is sufficiently different from the other signals. Another approach is to monitor SN on one or more subchannels by monitoring the "sync" frame.
To monitor R. The presence of jamming from data sources on adjacent telephone lines manifests itself as a change in subchannel SNR. A straightforward way of monitoring jamming events caused by activation or deactivation of devices that interfere with communication is that if any of these events occur,
The direct transmission of signals between the device and the local modem. As shown in FIG. 3, for example, the signal transmission line 3
5, 37 extend directly between the local modem 34 ′ and its associated device 31, 32, such as activating (“off hook”) or deactivating (“on hook”) these devices. Alternatively, changes in these devices may be directly transmitted as signals.

In addition to changing the control table in response to jamming events, to minimize interference to voice communications caused by upstream transmissions, and for those transmissions to leak into downstream signals ("echo"). It is desirable to reduce the upstream transmit power level in order to reduce These interferences result from non-linearities caused by devices such as phones that are coupled to the line, especially when the phone is off-hook.
The amount of power reduction required to make interference acceptable varies from phone to phone. In a preferred embodiment of the invention,
The probe signal is used to determine the required reduction in upstream transmit power. Specifically, after detecting an interfering event such as phone activation or deactivation, or interference from other sources that may disrupt communication, the ATU
-R transmitter section ("upstream transmitter") transmits a test signal over the subscriber line at varying power levels and echoes at the ATU-R receiver section ("downstream receiver"). To measure. The resulting measurement is the upstream transmit power level, which minimizes echo at the downstream receiver, or
Used to determine the upstream transmit power level that makes echo at least acceptable. The new power level is, of course, typically associated with the corresponding new parameter set of channel control parameters.

In addition to changing the bit allocation and gain parameters in response to a jamming event, changing one or both of the sub-channel equalizer (ie time domain equalizer or frequency domain equalizer) and echo canceller Is generally needed. A suitable set of these parameters can be preformed in the same manner as the bit allocation and channel gain (ie, in a pre-training session, with the various devices connected to the subscriber line activated, the line activated). Send a test communication via
The resulting communication conditions are measured and various parameters are determined based on the measurements) and can be stored in a secondary channel control table for recall or use as needed. Alternatively, the appropriate set of these parameters may be quickly re-determined during the retraining operation after detection of the jamming event without unduly disrupting communication. This is because these parameters are local to the receiver and therefore do not need to be transmitted to the other modem of the communication pair.

Specifically, in accordance with a preferred embodiment of the present invention, upon detecting a jamming event, the transceiver enters a "fast retrain" phase, as shown in more detail in FIG. A common jamming event is to take the phone off-hook or bring the phone back on-hook, which is commonly detected by ATU-R. Although the fast retraining process is shown for such events, the fast retraining process is not so limited and retraining is initiated at any end of the communication, for any type of disturbing event. It is understood to get. Therefore, when such an event (FIG. 8, event 200) is detected, ATU-
R notifies ATU-C to enter the fast retraining mode (step 202). The notification is preferably done by transmitting a specific tone to the ATU-C,
It may also include messages or other forms of communication. Upon receiving this notification (step 204), ATU-C
Is the A for the power level used for subsequent communications.
Wait for notification from TU-R. This includes at least the upstream power level and may also include the downstream power level.
This is because changing the upstream power level can affect downstream communications to some extent. For the sake of completeness it is assumed that both of these power levels are varied, but it is understood that in many cases only the upstream power level will be varied.

The new power level used is ATU-
Determined by R (step 208). ATU-R
Sends a channel probe test signal to the upstream transceiver and measures the resulting echo at the downstream receiver. The ATU-R then sets the upstream power level to minimize echo to the downstream signal, and also at the upstream transmitter to minimize the impact of upstream transmission leaking to downstream transmission. The downstream power level can be set. The ATU-R then selects the selected upstream, for example by transmitting one or more tones to the upstream transceiver, which are modulated by a binary PSK (Phase Shift Keying) signal to ensure robust transfer of power levels. And the downstream transmission level to ATU-C (steps 210, 212). The power level may be specified directly (eg as “-30 dbm”) or indirectly (eg as “level 3” for a given level group). This identification may identify the actual value of the power level, or simply the change in the power level that is implemented.

ATU-R (step 214) and AT
The UC (step 216) then begins transmission at the new power level for the purpose of retraining the equalizer and echo canceller. Preferably, the change to the new power level is synchronized by the use of a frame counter used in the DSL system to match the transmitter and receiver, but the synchronization is by other means (eg tone or By transmitting a message, or
It may be achieved (by simply sending a flag), or it may be left unsynchronized. Based on the training transmission, ATU-R and ATU-C determine appropriate time and frequency domain equalizer parameters for the new power level and appropriate echo canceller coefficients (steps 218, 220). The determination may include a calculation based on these measurements to determine the coefficient, the measurement from the one or more pre-calculated sets stored in the ATU-R and ATU-C respectively. May be used to select a particular set or sets of.

For example, the SNR on various subchannels as well as determining the power level in response to a jamming event.
Transmits a message or tone set to the other modem of the communication pair to identify the particular device or devices associated with the event, and thus, simply identify the number of parameter sets used for subsequent communication. Then, it can be used to select the appropriate pre-stored parameter sets stored in ATU-R and ATU-C respectively. Thus, the SNR measurement serves as the "signature" of the device or devices associated with the jamming event, and allows for rapid identification of these devices. This approach can significantly reduce the time required to retrain the equalizer and echo canceller. Then, even when training is required under a specific environment, the training time can be significantly reduced by using the coefficient stored in advance as the starting point.

To facilitate the use of SNR measurements in retrieving the corresponding parameter set, the various stored parameter sets are indexed into the set of SNRs to identify one or more associated with particular communication conditions. It is desirable to be able to quickly identify and retrieve a parameter set of One way this can be achieved is shown in FIG. 9A. Here, the first set 25
Each parameter set, such as 0, the second set 252, etc., is "on hook" (Table 250) in addition to the subchannel (SC) number 254 and corresponding bit allocation (BA) and gain (G) entries. , "Off-hook" (table 252), etc., specific SNR entries 26 for a parameter set suitable for a given communication condition.
Has 0. Additional parameter sets such as frequency domain equalizer coefficients, time domain equalizer coefficients, and echo cancellation coefficients may also be stored in the table, as appropriate for the receiver portion of the modem. For the transmitter part, these coefficients are not applicable,
Therefore, it is not stored.

Another means of linking the subchannel SNR and the corresponding parameter set is shown in FIG. 9B. As shown in FIG. 9B, a simple list structure 270.
Is a parameter set identifier 272 and a number of SNs.
R measurements 274, 276, etc. are included. The SNR for some or all of the subchannels may be included. The list may be the searched means for criteria to identify the closest match to the stored parameter set, which set is then retrieved and used thereafter. In either FIG. 9A or FIG. 9B,
The parameter set indexed to SNR is
It may be a set of multiple parameters, such as bit allocation and gain, among others, or it may include a single set, such as only bit allocation or only gain.

The identities of the channel control parameter sets used in subsequent communications are exchanged between transceivers (steps 226-232). These transceivers then switch to these parameter sets (2
34, 236), and start communication under the new conditions. The message containing the channel control parameters is preferably modulated in a similar manner to the "power level" message. That is, it is modulated with several modulation tones along with the BPSK signaling. Therefore, the message is short and very robust. Short messages are important so that fast retraining time is minimized. Because the modem is not transmitting or receiving data during this time, as is the case when the modem is used for eg video transmission or internet access, and therefore its temporary availability But,
Because it can be very noticeable. Similarly, it is important that the message transmission be robust. This is because error-free communication during jamming events is very difficult due to impulse noise, such as from reduced SNR, ringing or dialing. Therefore, the provision and use of a pre-stored set of parameters increases the reliability of the communication in spite of the absence of a splitter in at least one of the modems and in the presence of disturbing events that coincide with the data communication. Greatly increase.

It is anticipated that the modems described herein will most commonly be used in a dedicated pair. That is, each subscriber (ATU-R) modem communicates with a dedicated central office (ATU-C) modem. However, in certain cases, in order to service more than one subscriber modem,
It may be sufficient to provide a single master central office modem. The present invention also accommodates that possible event. Thus, in FIG. 10, the central office modem 280 has the switch 28
2 to communicate with a plurality of subscriber modems 284, 286, 288 via subscriber lines 290, 292, 294. These modems may be located at different distances from the central office and in different communication environments, so the channel control table of each modem is unique among itself. Therefore, there is one central office modem for each subscriber modem.
Store a master set 296 consisting of individual channel control parameter sets 298, 300, 302, etc. that are sets (both transmit and receive). When initiating a communication to a particular subscriber, the central office modem retrieves the appropriate transmission parameter set for the subscriber and uses it in subsequent communications. Similarly, when initiating communication to the central office, a given subscriber modem identifies itself and allows the central office modem to retrieve the appropriate set of receive parameters for that subscriber. To do.

Conclusion From the above it can be seen that the Applicant has provided an improved communication system for communication over a limited bandwidth subchannel such as a regular residential telephone line.
This system simultaneously accepts both voice and data communication over the line and eliminates the need for the installation and use of a "splitter." That is, it eliminates the expense that can impede the adoption and use of the high communication capabilities provided by DSL systems. Thus, this system can be implemented and used as widely as conventional modems are implemented and used today, but provides significantly greater bandwidth than is currently achievable with such modems. To do. [Brief description of drawings]

FIG. 1 is a block diagram of a conventional Digital Subscriber Line (DSL) system using a POTS splitter, typical of the prior art.

2 shows an exemplary bit allocation and gain table used in the apparatus of FIG.

FIG. 3 is a block diagram of a splitterless DSL system according to the present invention.

FIG. 4 is a block diagram of a splitterless transceiver according to the present invention.

FIG. 5A shows a channel control table constructed and used in accordance with the present invention.

FIG. 5B shows a channel control table constructed and used in accordance with the present invention.

FIG. 5C shows a channel control table constructed and used in accordance with the present invention.

FIG. 6 is a diagram of one form of a jamming event detector according to the present invention.

FIG. 7 illustrates the use of frame counters to communicate switching decisions to remote modems.

FIG. 8 illustrates a preferred procedure used to perform fast modem retraining in accordance with the present invention.

FIG. 9A illustrates how a channel control table can be easily selected in accordance with the present invention.

FIG. 9B illustrates how a channel control table can be easily selected in accordance with the present invention.

FIG. 10 shows another configuration of a modem interconnect of the present invention.

─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Klinsky, David M. USA Massachusetts 01720, Acton, Ayer Road 4 (72) Inventor Tazanes, Marcos USA Massachusetts 02173, Lexington, Unit No. 53, Lowell Street 665 ( 72) Inventor Tazanes, Michael A. USA Massachusetts 02173, Lexington, Carley Road 17 (56) References JP-A-11-168515 (JP, A) JP-A-60-112357 (JP, A) JP-A-9- 51328 (JP, A) JP 7-336274 (JP, A) JP 4-22235 (JP, A) JP 2000-514991 (JP, A) JP 62-502932 (JP, A) Europe Publication of patent application 820168 (EP, A 1) US Patent 5533008 (US, A) US Patent 5479447 (US, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04J 11/00

Claims (25)

(57) [Claims] 1. A modem for communicating data over a plurality of discrete sub-channels, each of the sub-channels
It is characterized by the bit allocation parameter which defines the bit assignments over the subchannel corresponding to communicate over the sub-channel, the allocating bits to said sub-channel during a first communication state means for storing the first channel control table, and means for storing the second channel control table allocating bits to said sub-channel during a second communication state, the second channel control table, disturbing event Equivalent to
And, the first and second channel control table is determined during initialization session communication capability of the sub-channel is determined, the first channel control table, determines a state where interference signaling condition does not exist Will be a modem. 2. The modem according to claim 1, wherein the second channel control table is determined as a function of the first channel control table. 3. The modem of claim 2, wherein the bit allocation of the second channel control table is determined as a percentage of the bit allocation of the first channel control table. 4. A modem for communicating data over a plurality of discrete sub-channels, each of the sub-channels
Is characterized by the bit allocation parameter which defines the bit assignments over the subchannel corresponding to communicate over the sub-channel, as the improvement, the sub-channel during a first communication state means for storing the first channel control table allocating bits to, and means for storing the second channel control table to assign the bits to the sub-channel during a second communication state, said second channel control Table for jamming event
Equivalent to, the second channel control table, a corresponding plurality of event multiple channel determined in response to a plurality of signaling events generated by the generation source
Is selected from the control table, each of the plurality of event generation source defines a certain channel control table at a given source, and means, when the source generates an event, the second channel system
For use as a control table, and means for selecting one of said plurality of channel control <br/> table, modem. 5. via the loop having both multiple upstream communication channels and downstream communication channels are formed from the sub-channel, a modem for use in asymmetric digital subscriber line communication, the
Loops can carry both voice and data communications
And defining a data communication between the modem and a second modem connected to the loop in a first communication state.
Means for storing a table, in the second communication state, defining the data communication between the modem and the second modem, to store the second table
The second table is a means for
Limit corresponding, and means, in order to investigate the transmission of return characteristics to said loop by said modem, and means for emitting a test signal in the loop, the output level of the transmission according to the investigated return characteristics and means for, the study includes a plurality of tones defined amplitude and frequency, the investigated return characteristics,
In response to emit said tone comprises at least one property selected from the group comprising the amplitude and frequency of the signal returned to the modem, and means, modem. 6. An assembly formed from a plurality of sub-channels.
Downstream communication channels and downstream communication
Asymmetric digital through loop with both channels
A modem used for Tal subscriber line communication,The
Loops can carry both voice and data communications
Is composed of Connected to the modem and the loop in the first communication state.
Defining a data communication with a second modem which is
Store a table ofDoMeans and The modem and the second modem in the second communication state
Stores a second table that defines data communication betweenYou
RumeansAnd the second table is
Corresponding means, Equalize transmission characteristics of the sub-channelDoEqualizerAnd
TheFirst and secondThe table is (1) Coefficient of time domain equalizer, (2) The coefficient of the frequency domain equalizer, (3) With the coefficient of the digital echo canceller Specify at least one ofEquipped with an equalizer
Themodem. 7. via the loop having both multiple upstream communication channels and downstream communication channels are formed from the sub-channel, a modem for use in asymmetric digital subscriber line communication, the
Loops can carry both voice and data communications
And defining a data communication between the modem and a second modem connected to the loop in a first communication state.
Means for storing a table of the first table, the first table being determined during an initialization process in the absence of the selected event, the modem and the second table in a second communication state. defining a data communication between the modem and to store the second table
The second table is a means for
A modem , which comprises means for hitting. Wherein said second table is determined during the initialization process in the presence of selected events,
The modem according to claim 7. 9. The modem of claim 8, wherein the second table is redetermined in response to the occurrence of a selected event. 10. Displaying the re-determined second table
10. The modem of claim 9, wherein the data is sent from a given modem to other modems with which it communicates in a dormant state. 11. A method of transmitting data over a wireline via upstream and downstream channels using a plurality of discrete frequency subchannels, first and second, wherein at least two different communications are provided. Storing, under conditions, at least first and second parameter sets that define data communication over the channel; and selecting a parameter set to be used in the communication, depending on the prevailing communication conditions. Transmitting a signal identifying the set of parameters to be selected over the line, the signal being transmitted on a sub-channel of an intermediate band between the upstream channel and the downstream channel. Be done. 12. A method of data transmission over a wireline via upstream and downstream channels using a plurality of discrete frequency subchannels, first and second, the method comprising at least two different communications. Storing, under conditions, at least first and second parameter sets that define data communication over the channel; and selecting a parameter set to be used in the communication depending on the prevailing communication conditions. Receiving a signal over the line identifying the parameter set to be selected, the signal being received on a sub-channel intermediate the upstream channel and the downstream channel. 13. A method of performing data transmission over a wireline via upstream and downstream channels from a plurality of discrete frequency subchannels, the first and second plurality of discrete frequency subchannels comprising at least two different communication states. Below, storing at least a first and a second parameter set defining a data communication over the channel, and selecting a parameter set to be used in the communication according to a prevailing communication condition. , The parameter set comprises at least one parameter set from the group comprising sub-channel frequency domain coefficients, time domain coefficients, and echo cancellation coefficients. 14. A method of communicating data over a plurality of discrete sub-channels, each of the sub-channels
It is characterized by the bit allocation parameter which defines the bit assignments over the subchannel corresponding to communicate over the sub-channel, the allocating bits to said sub-channel during a first communication state storing the first channel control table, it includes the step of storing the second channel control table in the second communication state allocating bits to said sub-channel, the channel control table of the second, interfering Equivalent to an event
And, the first and second channel control table is determined during initialization session communication capability of the sub-channel is determined, the first channel control table, determines a state where interference signaling condition does not exist Be done. 15. The second channel control table comprises:
15. The method of claim 14, determined as a function of the first channel control table. 16. The method of claim 15, wherein the bit allocation of the second channel control table is determined as a percentage of the bit allocation of the first channel control table. 17. A method of communicating data over a plurality of discrete sub-channels, each of the sub-channels
It is characterized by the bit allocation parameter which defines the bit assignments over the subchannel corresponding to communicate over the sub-channel, the allocating bits to said sub-channel during a first communication state storing the first channel control table, in the second communication state comprising the steps of storing the second channel control table allocating bits to said sub-channel, the second channel control table, disturbing event
Corresponds to, the second channel control table, a plurality of which is determined in response to a plurality of signaling events generated by the corresponding plurality of event generating source Cha
Is selected from the panel control table, each of the plurality of event generation source defines a certain channel control table at a given source, the steps, the source is an event
When generated, for use as the second channel control table, comprising a step of selecting one from the channel control table of the plurality of process. 18. A method of communicating data over a plurality of discrete sub-channels, each of the sub-channels
It is characterized by the bit allocation parameter which defines the bit assignments over the subchannel corresponding to communicate over the sub-channel, the allocating bits to said sub-channel during a first communication state step der storing storing the first channel control table, the second channel control table in the second communication state allocating bits to said sub-channel
Therefore, the second channel control table is
Corresponding to the step, when the first modem is in one of the communication states,
Check the channel of either the first or second channel control table.
Re-determining the channel control table and placing the sub-channel in communication with the first modem.
A second modem connected to Le, channel determined該再
Sending data representing a control table. 19. The method of claim 18, wherein the data representative of the re-determined channel control table is transmitted via a dedicated sub-channel selected from the discrete sub-channels. . 20. The method of claim 18, further comprising communicating information identifying a type of the re-determined channel control table to the second modem. 21. An array formed of a plurality of sub-channels.
Upstream communication channels and downstream communications.
An asymmetric digital signal through a loop that has both
A method for Tal subscriber line communication,The loop
Is configured to carry both voice and data communications
Has been done, Connects to the first modem and the loop in the first communication state.
Define data communication with a second modem that is continued,
Storing the first table, In the second communication state, the first modem and the second modem
A second table defining the data communication with the dem
Step to storeAnd the second table is
Steps corresponding to events, Equalizing the transmission characteristics of the sub-channelAnd
And the first and second tables are (1) Coefficient of time domain equalizer, (2) The coefficient of the frequency domain equalizer, (3) With the coefficient of the digital echo canceller Define at least one ofIncluding steps and
To doMethod. 22. A method for asymmetric digital subscriber line communication over a loop having both an upstream communication channel and a downstream communication channel formed from a plurality of sub-channels, the loop comprising:
Is configured to carry both voice and data communications
And defining a data communication between the first modem and a second modem connected to the loop in a first communication state,
Storing a first table, the first table being determined during an initialization process in the absence of the selected event, the first modem in steps and in a second communication state. Storing a second table defining a data communication between the second modem and the second modem, the second table comprising:
A method corresponding to an event, including steps . 23. The second table, the method according to determined, claim 22 during the initialization process in the presence of a selected event. 24. The method of claim 23, wherein the second table is redetermined in response to the occurrence of a selected event. 25. In hibernation, the redetermined second from a given modem to another modem with which it communicates .
25. The method of claim 24, further comprising transmitting data representing a table of